Preparation of Curcuma Longa L. Extract Nanoparticles Using Supercritical Solution Expansion

Journal of Pharmaceutical Sciences 108 (2019) 1581-1589

Contents lists available at ScienceDirect

Journal of Pharmaceutical Sciences journal homepage: www.jpharmsci.org

Pharmaceutical Nanotechnology

Preparation of Curcuma Longa L. Extract Nanoparticles Using Supercritical Solution Expansion Fatemeh Momenkiaei, Farhad Raofie* Department of Analytical and Pollutants Chemistry, Shahid Beheshti University, Tehran 1983963113, Iran

a r t i c l e i n f o

a b s t r a c t

Article history: Received 17 August 2018 Revised 25 October 2018 Accepted 7 November 2018 Available online 12 November 2018

Nanoparticles of Curcuma longa Linn (turmeric) rhizome extract were prepared using supercritical carbon dioxide (SC-CO2). The SC-CO2 was used for sample pre-treatment, including lipophilic compounds removal and extraction, as well as particle production. The particle formation process was based on the expansion of supercritical solution of plant extract into a secondary chamber. In the course of the expansion of supercritical solution process, the herbal extract changed from dissolved mode at higher pressures to precipitated mode at lower pressures, as long as the pressures were higher than the critical pressure. The particle growth via coagulation was limited by a large number of unsuccessful collisions between CO2 molecules and primary nuclei due to the presence of pressurized CO2 molecules where the particle formation occurs. The presence of curcumin derivatives in nanoparticles was confirmed by liquid chromatography-mass spectrometry results. Irregular to quasi-spherical particles with average diameter of 47 ± 20 nm (n ¼ 300) were prepared at a pre-expansion pressure of 35 MPa, pressure drop of 24 MPa, temperature of 50 C, equilibration time of 30 min, collection time of 60 min, extract volume of 30 mL, and feeding solution concentration of 2 mg mL1. © 2019 American Pharmacists Association®. Published by Elsevier Inc. All rights reserved.

Keywords: nanoparticle supercritical solution expansion Curcuma longa Linn curcumin particle design

Introduction Herbal medicines are gaining increasing deals of attention in the global market due to the side effects of synthetic medicine. Medicinal plants provide an excellent source of natural bioactives that are formulated by nature. This naturally formulated matrix is responsible for the stability of active herbal components, but suffers from limited bioavailability. A large number of the herbal medicines are administered orally in solid dosage forms such as tablets or capsules. However, this form of application is limited by limited bioavailability of most of the natural bioactive compounds as a result of poor solubility in aqueous media. Increasing the administration frequency or drug dose does not represent a good solution due to associated gastrointestinal disturbances, side effects, and decreased systemic drug clearance. Solubility of a powder is affected by various parameters such as size, size distribution, and morphology of its particles.1-3 In order to enhance bioavailability of such materials, several strategies have been undertaken, such as using advanced drug delivery systems. Among others, decreasing

This article contains supplementary material available from the authors by request or via the Internet at https://doi.org/10.1016/j.xphs.2018.11.010. * Correspondence to: Farhad Raofie (Telephone: þ98-02129903090). E-mail address: f_raofi[email protected] (F. Raofie).

the particle size (micronization methods) is one of the techniques of choice.4-9 Micronization techniques are used to improve dissolution rate of a solute into biological environments by increasing contact area.4,5 It is also expected that an active material can blend with an excipient more uniformly when its particles are smaller and exhibit narrower size distribution. Nowadays, food and pharmaceutical industries use numerous methods for preparing materials with smaller particles. These methods include top-down strategies such as mechanical comminution (e.g., crushing, grinding, and milling), and bottom-up strategies such as spray-drying, freeze-drying, and anti-solvent re-crystallization. Thermal and chemical degradation of products, high consumption of organic solvent, broad particle size distribution, polymorph diversity, and solvent residue are major problems faced in conventional micronization methods.4,6 Supercritical fluids, especially supercritical carbon dioxide (SCCO2) (Tc ¼ 31.8 C, Pc ¼ 7.39 MPa), have been widely used as a medium (either solvent or anti-solvent) for submicron and nanoscale material fabrication thanks to their relatively low critical temperature and moderate critical pressure.6 In addition, the fluids are inexpensive, non-toxic, non-flammable, and can be easily evaporated from the remaining substance. The most important advantages of using the SC-CO2 technology include high purity of products, simplicity of the associated processes, possibility of producing solid particles with unique morphologies, mildness of

https://doi.org/10.1016/j.xphs.2018.11.010 0022-3549/© 2019 American Pharmacists Association®. Published by Elsevier Inc. All rights reserved.

1582

F. Momenkiaei, F. Raofie / Journal of Pharmaceutical Sciences 108 (2019) 1581-1589

operating conditions, possibility of obtaining fine particles with narrow size distribution, and the wide range of achievable materials. Several methods for small particle formation based on SC-CO2 techniques have been reported, including crystallization from supercritical solution, rapid expansion of supercritical solutions (RESS), supercritical anti-solvent process, solution-enhanced dispersion by supercritical fluids, aerosol solvent extraction system, gas anti-solvent, and particles from gas-saturated solutions/ suspension.5-9 In the RESS process, the SC-CO2 (or modified SC-CO2), containing the substance to be micronized, is subjected to rapid expansion through a heated nozzle. The degree of supersaturation is proportional to the concentration of the solute in supercritical solution, which directly depends on pre-expansion and post-expansion pressures and temperatures. A sudden pressure drop decreases the fluid density and hence solvating power of CO2. Faster pressure drops or higher pre-expansion pressures lead to an increase in the number of initial nuclei. There are 2 competitive nuclei-growing approaches, namely coagulation and condensation that are driven by collision and deposition, respectively.10-12 Conventional RESS shows such benefits as low solvent consumption and easy procedure. However, due to uncontrollable expansion, it suffers from broader particle size distribution compared with other supercritical fluidebased material processing techniques. In order to overcome associated drawbacks with the RESS (especially for the RESS with co-solvent; e.g., broad particle size distribution), a number of attempts have been undertaken. As an example, Pessi et al. developed a particle production method based on the expansion of supercritical solution (ESS). In the presented method (controlled expansion of supercritical solution (CESS)), expansion was practiced under mild, steady, and isothermal conditions. In the CESS, the post-expansion pressure was much higher than that in the RESS (which is equal to atmospheric or nearatmospheric pressure) at the time of expansion. Through the CESS, they collected small particles with a narrow size distribution.13,14 Curcuma longa Linn (turmeric) is one of the most popular spices in the world. As of now, more than 200 different bioactive compounds have been identified in turmeric, including phenolic compounds, sesquiterpenes, monoterpenes, sterols, alkaloids, curcuminoids, and essential oils.15,16 The presence of lipophilic compounds in turmeric has been confirmed by several research works.17-19 Curcuminoids (i.e., curcumin, demethoxycurcumin, and bisdemethoxycurcumin) are the main constituents of C longa Linn (turmeric) rhizome, an important natural pigment widely used as a spice and coloring agent in food products. The turmeric rhizome possesses potent antioxidant, anti-inflammatory, anti-diabetic, anti-Alzheimer, anti-microbial, and anti-cancer constituents, along with promoting agents.20-25 These extractable bioactives comprise 9%-60% (by weight) of turmeric rhizomes with curcuminoids (0.4%33%) as principal compounds.26-34 The yield of total extractable compounds (0.5%-2%; by weight) and curcuminoid contents (0.3%4%) were reported using SC-CO2 extraction with different co-solvents.28-34 Similar to a large number of important food and pharmaceutical active ingredients, curcumin is hardly water-soluble, diminishing its bioavailability in aqueous media and body liquids after oral administration.35,36 A substantial deal of research has focused on the enhancement of turmeric extract absorption.24,37-44 Many of these works were conducted to prepare amorphous turmeric extract nanoparticles with pharmaceutically approved excipients.24,39-44 This study is motivated by our previous work,45 in which herbal extract nanoparticles (Silybum marianum seeds extract) were prepared using the ESS. We purposed the possibility of nanoparticles

production from turmeric rhizome extract using the developed method. In addition, evaluation of the parameters affecting the particle characteristics in this experimental condition was aimed. Materials and Methods Reagents HPLC grade acetonitrile and ethanol (EtOH) were purchased from Caledon (Georgetown, ON, Canada). Dimethylformamide (DMF) was purchased from Merck & Co. (Merck KGaA). All other compounds used in this study were of analytical grade. The formic acid used as additive for liquid chromatography-mass spectrometry (LC-MS) eluent was supplied by Fluka (distributed by SigmaAldrich, Allentown, PA). All aqueous solutions were prepared using HPLC grade water (Merck KGaA). Carbon dioxide (99.99% purity), contained in a cylinder with an eductor tube, was obtained from Roham Co. (Tehran, Iran). Plant Material and Sample Preparation C longa Linn (turmeric) rhizomes were collected from Zahedan (Iran) in August 2017. Those were manually cleaned to separate extraneous matters such as dust, and then stored in a dark place and dried at room temperature for 1 month. The rhizomes were milled using mortar and pestle, crushed by a laboratory mill (Myson, China), and passed through a set of standard mesh sieves (Azmoon Saz Mabna Co., Tehran, Iran). Finally, a powder with particle size of about 0.5 mm was collected and stored in dark at 5 C ± 1 C. Equipment for Supercritical Fluid Experiments Experimental works, including the clean-up, extraction, and particle collection, were carried out using a laboratory scale supercritical fluid extraction (SFE) system (Suprex MPS/225 System; Suprex Corp., Pittsburgh, PA) (Fig. 1). The instrument was adjusted on the SFE mode a maximal operating pressure of 39.5 MPa. A Dura-flow manual variable restrictor (Suprex) on the SFE system was electrically heated to avoid freeze-plugging during expansion of the SC-CO2. Figure 1a shows a schematic diagram of the apparatus used for SC-CO2 extraction and clean-up. Pure CO2 gas was supplied from a cylinder (1), liquefied by a condenser (2), and subsequently pumped by a high-pressure syringe pump unit (3). CO2 was heated using a heat exchanger (4) and passed into the oven (5) before entering the 5-mL stainless steel vessel (extraction cell, 6). Both sides of the extraction cell were equipped with a pair of sinter metal filters (with a pore size of 0.1 mm). The pressure was controlled by a back-pressure regulator. The restrictor was heated at 100 C (7). During dynamic extraction through the restrictor, SC-CO2 flow rate was approximately 0.30 ± 0.05 mL min1. The instrumental setup used for particle production is depicted in Figure 1b. A 25-mL stainless steel vessel (8) was coupled with the extraction cell (6) via a needle valve (valve 3). Two sintered metal filters (with a pore size of 0.1 mm) were attached to each end of the equilibration and collection chambers. The instrument setup was quite similar to the SFE setup (Fig. 1a), except that a 25-mL collection chamber (8) was coupled with the extraction chamber via a needle valve (valve 3). Therefore, the setup consisted of 2 main units: equilibration (6) and collection chambers (8). Supercritical Fluid Pre-Treatment and Extraction Plant powder (1.0 g) was well mixed with 1-mm diameter glass beads and charged into the extraction vessel. The first step of the

F. Momenkiaei, F. Raofie / Journal of Pharmaceutical Sciences 108 (2019) 1581-1589

1583

Figure 1. Schematic instrument setup designed for (a) supercritical clean-up and extraction and (b) particle production process.

process was performed to clean-up the sample from lipophilic compounds using pure SC-CO2. The clean-up stage was carried out at a temperature of 40 C, a pressure of 35 MPa, and static and dynamic extraction times of 30 and 60 min, respectively. The extracted material was collected into a 3-mL vial containing 1 mL of EtOH. The collected sample was placed in dark at room temperature for 2 weeks, until dryness. The mass of extracted lipophilic compounds was determined gravimetrically. The extraction yield was calculated as the ratio of the dry weight of extract to the weight of seeds  100 per g of seed. The experiment was followed by a second step involving the extraction of the active substance upon spiking 160 mL of EtOH as a modifier, which was carried out by SC-CO2 extraction at 40 C, 34 MPa, 30 and 70 min of static and dynamic extraction times, respectively. The extract was collected in EtOH in a 5-mL vial, while the flow rate of SC-CO2 was about 0.30 ± 0.05 mL min1 during dynamic extraction time. To improve the collection efficiency, the volumetric flask was placed in an ice bath during the dynamic extraction stage. For all the extraction trails, the modifier (EtOH) was spiked directly into the sample in the extraction vessel before attaching the extraction system to the SFE system. The extraction yield (herbal extract containing curcumin) was calculated by weighing the collected substance after evaporation of EtOH at 40 C at the end of each run. Concentration of the extracted solution was adjusted to 2 mg mL1 of dried herbal extract in EtOH. The resultant solution was used for particle production. For this purpose, 100 mL of the resultant solution was spilled into a 1-mL screw conical-bottomed glass test tube and evaporated to dryness at 40 C (TSa). Prior to further use, the 1mL screw conical-bottomed glass test tubes were placed in a furnace at 500 C for 3 h to remove any organic pollutants. The vial was capped and stored at 5 C ± 1 C until LC-MS analysis. The extract solution was used for either the particle production step or other applications, and stored in dark at 5 C ± 1 C. Supercritical Particle Collection In this step, 2 stainless steel vessels were coupled with a valve between them (Fig. 1b). The first and the second vessels were 5 and 25-mL extraction cells, respectively. The first cell was used as an equilibration chamber and the second 1 served as the collection chamber. Both sides of the chambers (equilibration and collection chambers) were equipped with stainless steel filters (with a pore size of 0.1 mm). With an oriflamme on its wall, a 500-mL polyethylene capsule containing 20-40 mL of 1 or 2 mg mL-1 herbal extract in EtOH solution was placed in the first extraction vessel. The oriflamme was big enough for SC-CO2 diffusion into the capsule, while it was small

enough to allow the liquid solution to come out. Mica surfaces (Agar Scientific Ltd.) were used as clean imaging substrates for particle collection. The mica sheets were cut into 5  50 mm rectangular pieces using sharp scissors and freshly cleaved just before use. The collecting surfaces (3 mica sheets) were placed in the collection chamber. The mica sheets remained stable during the experiment. Cu grids were placed in the collection chamber for high-resolution transmission electron microscope (HRTEM) measurement (experiments 12 and 13). Outlet tube of the restrictor was immersed in a vessel containing 50 mL of DMF. In order to ensure complete collection of the solvent (20-40 mL of EtOH), the collection vessel (solvent trap) was placed in an ice bath. A schematic diagram of approximate pressure profile and system configurations (valve assembly) during the particle collection is presented in Figure 2a and Table 1, respectively. The system was loaded at room temperature and atmospheric pressure with the capsule containing extract (20-40 mL) and mica sheets, respectively (Fig. 2a, point 0; Table 1, row 0). The particle collection was carried out at different constant temperatures (i.e., 40 C, 45 C, 50 C, or 60 C, depending on the setpoint value). The system was pre-filled with SC-CO2 at a pressure of 8 MPa for about 1 min (Fig. 2a, point 1; Table 1, row 1), then valve 3 was closed, and pressure of the equilibration chamber was increased to 35 MPa (Fig. 2a, point 3; Table 1, row 2) for 10-30 min. During the equilibration time (Fig. 2a, points 3 and 4; Table 1, row 3), available free space within the first vessel was saturated with the extracted material dissolved in supercritical CO2 modified with EtOH (SC-CO2-EtOH) solution. Once the equilibrating time (10-30 min) was passed, valve 3 was opened while valve 4 was also opened at a flow rate of 0.1 mL min1 until the post-expansion pressure was reached (Fig. 2a, points 4 and 5; Table 1, row 4). During the precipitation period, valve 4 was closed for 60 min while the instrument was programmed at the post-expansion pressure and constant operating temperature (Fig. 2a, points 5 and 6; Table 1, row 5). Subsequently, valve 4 was adjusted to a flow rate of 0.5 mL min1 for 3 h, allowing 90 mL of SC-CO2 (at a constant pressure of P2 MPa) to be purged through the system to remove the residual EtOH (Fig. 2a, points 6 and 7; Table 1, row 6). At the end of the experiment, valve 2 was closed, the precipitation vessel was de-pressurized through valve 4 (Fig. 2a, points 7 and 8; Table 1, rows 7 and 8), and the solvent collection vessel was capped and stored at 5 C ± 1 C until gas chromatography (GC) analysis (TSb; collected during experiment 14). Then, the collection chamber was disclosed and a small piece of collective surfaces (mica sheets) was placed in the closed vessel until the particle size determination was accomplished. The rest of the sheets were cut into small pieces and soaked into a 5-mL vial containing 3 mL of EtOH. In each experiment, 2 sheets of mica (equivalent to a total

1584

F. Momenkiaei, F. Raofie / Journal of Pharmaceutical Sciences 108 (2019) 1581-1589

Figure 2. The approximate pressure profile schematic diagram used in (a) this study, (b) RESS, and (c) CESS. In diagram (a), the pressure profiles within the equilibration and collection chamber are indicated by a straight and dashed line, respectively. The diagrams (b and c) are reported with reference to Pessi et al.’s13 study.

surface area of 500 mm2 at both faces of the sheet) were washed. The mica sheets were selected from 9 experiments (equal to 4500 mm2). The solution was bubbled to dryness with CO2 gas at 40 C. The vial was capped and stored at 5 C ± 1 C until LC-MS analysis (TSc). Characterization of Nanoparticles and Extract Elemental analysis of the collected nanoparticles was performed using energy dispersive X-ray (EDX) spectroscopy (VEGAII

TESCAN). The characterizations were carried out on the extract (SFE sample; TSa) and nanoparticles (TSc) using LC-MS, according to the report by Jiang et al.46 The analysis was performed using an LC-MS system manufactured by Agilent Technologies (Waldbronn, Germany). The system was made up of an Agilent 1200 LC system coupled with an Agilent 6410 triple quadruple tandem mass spectrometer. Data acquisition and instrument control were provided using Agilent Mass Hunter Workstation Software. Analytical separation was performed using an Agilent Zorbax Stable Bond analytical column (Agilent HT Zorbax SB-C18 column; 1.8 mm, 2.1 mm  50 mm; Agilent Technologies, Santa Clara, CA). Binary gradient elution was carried out using an aqueous solution of 0.1% formic acid (mobile phase A) and a mixture of mobile phase A/acetonitrile (30:70; mobile phase B). The elusion was started at 10% mobile phase B, changed to 90% in 20 min, and held for 25 min. The flow rate and the injection volume were 0.25 mL min1 and 5 mL, respectively. The electrospray ionization voltage was fixed at 4 kV (in positive ion mode). The ultra-high purity nitrogen was used as the nebulizer and collision gas. The capillary temperature was maintained at 300 C after the initial heating step. Acted as drying gas, the nebulizer gas was adjusted at the flow rate and maintained the pressure of 10 L min1 and 0.27 MPa, respectively. In the next step, 100 mL of 50% eluent A/50% eluent B was added to each of the test tubes containing the dried extract residues (either TSa or TSc), and the tube was shaken well until a uniform solution was obtained. Finally, 5 mL of the resulting solution was injected into the LC-MS system. Solvent residues were measured using a gas chromatograph equipped with a flame ionization detector (GC-FID; Varian CP-3800 system equipped with a DB 624; Agilent fused-silica capillary column, 30 m  0.450 mm; film thickness: 2.55 mm). Residual solvent determination was carried out according to the study by Witschi and Doelker,47 as described in the following: first, column temperature was maintained at 30 C for 5 min, then raised to 230 C at 10 C min1 and maintained there for 10 min. The injection port and detector were maintained at 140 C and 250 C, respectively. Nitrogen at a pressure of 0.027 MPa was used as carrier gas. Particle Size and Size Distribution Determination Size and size distribution of the collected particles were characterized by field emission scanning electron microscopy (MIRA3 TESCAN-XMU, Brno, Czech Republic). Image analysis was performed using Microstructure Measurement software. Output data from the image analysis were assembled in Microsoft Excel. Results of the particle size analysis were further confirmed by atomic force microscopy (AFM) measurements conducted on a multimode

Table 1 System Configuration During the Particle Production Process Stage

0 1 2 3 4 5 6 7 8

Valve

Pressure of Chambers (MPa)

1

2

3

4

Equilibration

Collection

þ þ þ þ þ þ þ þ þ

þ þ þ þ þ þ þ  

þ þ   þ þ þ þ þ

    þ  þ þ þ

Atmospheric 8 P1 P1 P1 P2 P2
Atmospheric 8 8 8 8 to P1 P2 P2
Different stages (1-8) are depicted in Figure 2a. þ, valve is open; , valve is closed; P1, pre-expansion pressure; P2, post-expansion pressure.

Status

Start-up Start-up Start of pre-expansion condition Pre-expansion Start to expansion Post-expansion System flushing Start of depression Depression

F. Momenkiaei, F. Raofie / Journal of Pharmaceutical Sciences 108 (2019) 1581-1589

nanoscope DPN 5000, a digital instrument under ambient condition. Morphology of the collected particles was investigated by an HRTEM (JEOL 2000). The turmeric extract nanoparticles were collected on carbon-coated Cu grids placed in the collection chamber (Fig. 1b). Plant Powder Thermal Analysis Differential scanning calorimetry (DSC) analysis was carried out on a Mettler TA 4000 instrument (Mettler Toledo, Greifensee, Switzerland). Accurately weighted plant powder (3 mg) was scanned at a heating rate of 10 C min1 over the temperature range of 25 C-200 C. Experimental Design Methodology In the present work, the data analysis and the design of the experimental matrix were carried out by the aid of STATGRAPHICS statistical software. Demonstrating the relationships between variables and responses graphically, the precise optimum point could be found by the aid of response surface methodologies. Results and Discussion Supercritical Fluid Pre-Treatment (Clean-Up) The presence of lipophilic compounds “extractable with SCCO2” in turmeric has been frequently reported in the literature.17-19 The lipophilic compounds can interfere with fine particle collection; as such, those shall be removed. Optimum conditions for the clean-up step were determined through central composite design (Supplementary Information is presented in S3.1, Tables S1-S3 and Fig. S1). An extraction yield of 0.02%-0.04% (by weight) was obtained for the extracted lipophilic compounds where the experiment was carried out at the optimum condition (a pressure of 35 MPa, a temperature of 40 C, and static and dynamic extraction times of 30 and 60 min, respectively). The result was within the range of previously reported values.17,28 The extracted oil was pale yellow in color, while the color of the de-oiled residue was quite yellowish due to the presence of curcuminoids. Supercritical Fluid Extraction The choice of EtOH as co-solvent was based on preliminary literature data demonstrating that EtOH extract exhibits more phenolic content and antioxidant activity.31 Ashraf et al.31 found that, in terms of polarity, EOH is more compatible with turmeric polyphenols, as compared to acetone and methanol. EtOH has been frequently used as a solvent in SC-CO2 particle production due to its sufficient miscibility with SC-CO2.44 It has been also used as cosolvent for SC-CO2 extraction. Optimum conditions for SFE were determined by response surface methodology: pressure, temperature, modifier volume, static and dynamic extraction times of 34.4 MPa, 40 C, 158.7 mL, 10 min, and 70.2 min, respectively (Supplementary Information can be found in S3.2, Tables S4-S6 and Fig. S2). The yield obtained at the selected conditions was found to be 2.23%. The result was in reasonable agreement with literature.17,26,28 Results of Characterization of Nanoparticles and Extract Figure S3 shows the EDX spectra of the collected particles on the mica sheets. The presence of carbon atoms coming from turmeric extract was confirmed by the EDX spectrum. The peaks correspond to Al, Si, and K, and fractions of oxygen come from mica sheet.

1585

Table 2 Multiple Reaction Monitoring Transitions and the Corresponding m/z Values for All the Expected Curcuminoid Fragments Variable

Bisdemethoxycurcumin

Demethoxycurcumin

Curcumin

RT in TSa (min) RT in TSc (min)

21.0 21.1

20.8 20.9

20.6 20.7

Fragment

m/z

m/z

m/z

C21H21O6 C21H19O5 C20H19O5 C20H17O4 C19H17O4 C18H19O4 C19H15O4 C17H17O4 C17H17O3 C15H15O4 C16H15O3 C14H13O4 C16H15O2 C14H13O3 C15H13O2 C13H11O3 C11H9O3 C10H9O3 C11H11O2 C9H7O2 C10H9O

e e e e 309a e 291 e e e e e 239 e 225a 215 189a e e 147a 145b

e e 339a 321a e e e e 269b e 255a 245a e 229a e e e 177a 175a 147a e

369a 351 e e e 299 e 285b e 259a e 245a e e e e e 177a 175a e e

RT, retention time. a Fragment was detected in the extract (TSa). b Fragment was detected in both the extract and the nanoparticles (TSa and TSc).

The LC-MS analysis was performed on the turmeric extract (TSa) and nanoparticles (TSc), so as to identify curcuminoids. The m/z value of the most abundant curcuminoid fragment ions are summarized in Table 2. Moreover, the identified fragment ions in SFE sample and nanoparticles are listed in Table 2. Total ion chromatogram of the nanoparticles (TSc) is presented in Figure S4a. The extracted ion chromatograms at m/z values of 369, 339, and 309 confirmed the presence of curcumin, demethoxycurcumin, and bisdemethoxycurcumin at retention times of 21.1, 20.9, and 20.7, respectively (Figs. S4b-S4d). MS spectra of the peaks at retention times of 21.1 and 20.9 min are depicted in Figures S4e and S4f, respectively. In a similar way, 3 curcumin derivatives were identified in turmeric extract (TSa). Accordingly, curcumin, demethoxycurcumin, and bisdemethoxycurcumin were identified in SFE sample using the corresponding precursor ions at m/z values of 369, 339, and 309 ([MeH]þ) at retention times of 21.0, 20.8, and 20.6 min, respectively. The curcumin content of 4.53%, 6.84%, and 6.89% was obtained for the plant powder, the extract, and nanoparticles, respectively. The LC-MS result demonstrated that the composition of curcuminoids in nanoparticles was 8.7%, 21.1%, and 70.2% for bisdemethoxycurcumin, demethoxycurcumin, and curcumin, respectively (Supplementary Information S3.3 and Figs. S5a-S5c). In order to quantitatively verify the capability of the method for removing residual solvents, the content of EtOH in TSb was measured by GCFID. The EtOH content in TSb (collected during experiment 14) was compared with the reference solution (30 mL of EtOH in 50 mL of DMF). The result demonstrated that the co-solvent can be completely resolved from the inner parts of SFE system at the optimized experimental conditions. Particle Collection Detailed experimental conditions and the results obtained from the particle production experiments are summarized in Table 3.

1586

F. Momenkiaei, F. Raofie / Journal of Pharmaceutical Sciences 108 (2019) 1581-1589

Table 3 The Experimental Conditions for Particle Production From Turmeric Extract Using SC-CO2 Number

P1

P2

dP

T

t1

t2

V

C

1 2 3 4 5 6 7 8 9 10 11

35 35 35 25 35 35 35 35 35 35 35

10 10 10 10 10 11 12 11 11 11 11

25 25 25 15 25 24 23 24 24 24 24

50 50 50 50 50 50 50 40 45 60 50

30 30 30 30 30 30 30 30 30 30 10

60 60 60 60 60 60 60 60 60 60 60

20 30 40 30 30 30 30 30 30 30 30

2 2 2 2 2 2 2 2 2 2 2

Observations

e e

12 13 14

35 35 35

11 11 11

24 24 24

50 50 50

30 30 30

60 60 60

30 30 30

1 2 2

Figures 4a and 4b e Figures 3c and 3e

e e e e e Figures 3b and 3d e e

The whole solvent was removed from the extraction cell and capsule The observation was the same as in experiment 1 Extra amount of EtOH remained within the capsule No particle was formed Instrument failed to work properly Irregular nanoparticles with a size of 47 ± 18 nm were obtained A degree of aggregation was observed A small amount of particle coalescence or aggregation was detected Irregular nanoparticles with a size of 58 ± 19 nm were obtained Irregular nanoparticles with a size of 60 ± 21 nm were obtained Small amount of feeding solution (turmeric extract) remained within the capsule Fewer particles were detected Irregular nanoparticles with a size of 48 ± 22 nm were obtained Irregular nanoparticles with a size of 45 ± 19 nm were obtained

SC-CO2, supercritical carbon dioxide; P1, pre-expansion pressure (equilibration pressure; 25 and 35 MPa); P2, post-expansion pressure (collection pressure; 10, 11, and 12 MPa); dP, pressure drop (MPa); T, temperature (40 C, 45 C, 50 C, and 60 C); t1, equilibration time (10 and 30 min); t2, particle collection time (60 min); V, volume of turmeric extract solution (feeding solution) or co-solvent (20, 30, 40 mL); C, concentration of turmeric extract solution (feeding solution; 1 and 2 mg mL1).

Based on the preliminary studies and experiments, 7 parameters were taken into consideration: pre-expansion pressure (equilibration pressure) (P1; 25 and 35 MPa), post-expansion pressure (collection pressure) (P2; 10, 11, and 12 MPa), temperature (T; 40 C, 45 C, 50 C, and 60 C), feeding solution volume (extract) (V; 20, 30, and 40 mL), equilibration time (t1; 10 and 30 min), particle collection time (t2; 60 min), and feeding solution concentration (C; 2 and 1 mg mL1 of undiluted or 2-fold diluted primary extract, respectively). The parameters were defined at different levels (T, P2, V, P1, t1, C, and t2 at 4, 3, 3, 2, 2, 2, and 1 levels, respectively). The effect of each parameter was evaluated by varying 1 parameter at a time, while the other parameters were held constant. Three of the 12 experiments led to successful formation of nanoparticles (repeats are not considered). Figure 3b shows the field emission scanning electron microscopy image obtained from experiment 6, indicating that the collected nanoparticles were uniformly dispersed in shape and size; moreover, the size distribution was narrow and a very small degree of aggregation was observed. Average particle diameter was 47 ± 18 nm (n ¼ 300; Fig. 3d). The nanoparticles were significantly smaller than those in the unprocessed air-dried extract (Fig. 3a). The result obtained for the 2 replicates of experiment 6 (experiments 13 and 14) demonstrated that the particle production process was reproducible. Average particle diameters were 48 ± 22 nm (n ¼ 300) and 45 ± 19 nm (n ¼ 300; Figs. 3c and 3e) for experiments 13 and 14, respectively. The collected nanoparticles exhibited an irregular spherical morphology. This is in good agreement with the literature24,39-44 and confirmed by the HRTEM results. Figure 4a shows the HRTEM image of the nanoparticles (experiment 12) deposited on the Cu grid and demonstrates that the collected particles are quasispherical. Detailed characterization of the nanoparticles collected in the same experiment (experiment 12) was carried out using the AFM measurements (Fig. 4b). The existence of the aggregation was confirmed by the AFM measurement indicating the nanoparticles were aggregated; however, the size of aggregates is smaller than 240 nm. In Vitro Dissolution Study The result of in vitro dissolution study showed an enhanced dissolution rate for nanoparticles compared to the air-dried extract. Nanoparticles exhibit dissimilar dissolution profile, as determined by statistical analysis of the data (f1 and f2 values), to those of the extract. Dissolution profiles of nanoparticles and the extract were

considered different because the calculated factors f1 (25) and f2 (35) did not meet acceptance criteria (0  f1  15; 50  f2  100). According to the United States Pharmacopeia monograph for curcuminoid tablet and capsule, the threshold limit for dissolved curcumin at 60 min is 75% (not less than 75.0%). The amount of dissolved curcumin at 60 min for the air-dried extract was close to the lower limit (80%), while it is significantly upper than lower threshold limit for nanoparticles (93%) (Supplementary Information S3.3, Table S7, and Figs. S5 and S6). Definition of Working Range for the Selected Parameters The ratio of co-solvent to SC-CO2 was determined by the feeding solution volume. Higher amount of co-solvent was desired to achieve better solubility of the considered compounds in EtOH. On the contrary, there is a limiting value for the amount of co-solvent beyond which the complex mixture of turmeric extract is likely to be fractionated. In order to avoid the fractionation of the extract, the entire volume of the feeding solution was aimed to be dissolved in 5 mL of SC-CO2 at the pre-expansion condition. An extra amount of solution remained within the capsule when the capsule was charged with 40 mL of feeding solution (experiment 3). The results obtained from experiments 1 and 2 revealed that the co-solvent volume of 30 mL was suitable. The selected value was large enough to improve the solubility of SC-CO2 (5 mL at the pre-expansion condition) toward the complex mixture of turmeric extract, while it is small enough to be dissolved all at once to prevent fractionation. The amount of substance to be precipitated is directly dependent on the pre-expansion pressure, extract concentration, and volume. Concentration of the extract (2 mg mL1) was held next to the highest value (2.6 mg mL1). The density and hence solvating power of SC-CO2 toward the considered component of the herbal extract increased at higher pre-expansion pressures. This was confirmed by the results obtained through experiment 4, where a few numbers of particles were formed at the pre-expansion pressure of 25 MPa. Effective precipitation of the dissolved turmeric extract during the post-expansion conditions was aimed. The precipitation of EtOH extract of turmeric using supercritical anti-solvent experiments has been reported previously. The SC-CO2-EtOH mixture is considered as precipitation medium for turmeric extract at pressures between 10.0 and 15 MPa.44 On this basis, the post-expansion pressure was investigated within the range of 10-12 MPa (experiments 5, 6, and 7). The post-expansion pressure of 11 MPa was selected as optimum value. Either the quantity of dissolved material

F. Momenkiaei, F. Raofie / Journal of Pharmaceutical Sciences 108 (2019) 1581-1589

1587

Figure 3. Field emission scanning electron microscopy images of turmeric extract nanoparticles: (a) unprocessed (size bar: 1 mm) and (b and c) experiments 6 and 14, respectively (size bar: 200 nm). The particle size distribution diagrams and average particle diameters (APD) correspond to (d) experiment 6 and (e) experiment 14.

or the amount of precipitate is determined by the capacity of the supercritical solution at pre-expansion or post-expansion conditions, respectively. A homogeneous SC-CO2-EtOH solution containing turmeric extract was prepared during equilibration time (t1 ¼30 min). An extra amount of solution remained within the capsule when the equilibration time was lowered from 30 min to 10 min (experiment 11). Therefore, the equilibration time of 30 min was approved for the rest of the experiment.

A DSC measurement was performed on the turmeric powder. The DSC baseline remained stable over the entire temperature range from 25 C to 105 C indicating that the plant powder is chemically stable below 105 C (Fig. S7). Therefore, the working range for temperature was selected between 45 C and 60 C. However, lower temperatures were focused preferably. Effect of temperature was investigated at different temperatures, namely 40 C, 45 C, 50 C, and 60 C, through experiments 8, 9, 6, and 10, respectively (the rest of the parameters were held constant at

Figure 4. Turmeric extract nanoparticles collected during experiment 12: (a) HRTEM image (size bar 100 nm) and (b) AFM image.

1588

F. Momenkiaei, F. Raofie / Journal of Pharmaceutical Sciences 108 (2019) 1581-1589

P1 ¼ 35 MPa, P2 ¼ 11 MPa, t1 ¼ 30 min, t2 ¼ 60 min, V ¼ 30 mL, C ¼ 2 mg mL1). Some sort of aggregation was detected at temperatures below 45 C (experiments 8 and 9) due to the coalescence of particles or the re-dissolution in the collection chamber. Nanoparticles were significantly less aggregated and well dispersed at 50 C and 60 C (experiments 6 and 10, respectively). Similar to our previous study,45 the particle production process consisted of 2 steps: equilibration and precipitation. A homogeneous SC-CO2-EtOH solution containing turmeric extract was prepared during equilibration time at higher pressures. Then, the turmeric extract-saturated supercritical solution was expanded into the collection chamber through a heated needle valve. At the time of expansion, the equilibration chamber is filled with supercritical solution (EtOH-SC-CO2 containing the extract) at the pressure of 35 MPa, while the collection chamber is filled with SC-CO2 at the pressure of 8 MPa. When valve 3 is opened, the supercritical solution leaves the equilibration chamber and enters to the collection chamber through valve 3. The valve itself cannot control the speed of the fluids. The speed of the supercritical solution is limited by the resistancepressurized CO2 molecules. For this reason, the expansion occurs in a controlled manner (instead of rapid manner). Here, compared to RESS (Fig. 2b), expansion occurred at a lower velocity. The pressure profile used in this study (Fig. 2a) was basically similar to that of CESS (Fig. 2c), except for the postexpansion pressure which was much higher than those used by Pessi et al.13 in CESS experiment. A large number of collisions occurred between CO2 molecules and initial nuclei because the system is pre-filled with SC-CO2. Therefore, the particle growth caused by coagulation was less efficient. Conclusion SC-CO2 extract obtained from de-oiled C longa Linn (turmeric) rhizome was converted to nanoparticles. Subsequent steps including clean-up, extraction, and particle production were carried out using SC-CO2. The particle production was based on the expansion of the supercritical solution. The process was designed to remove residual organic solvent, so as to reach the specified limit. Irregular to quasi-spherical particles with average diameter of 47 ± 20 nm (n ¼ 300) were prepared at a pre-expansion pressure of 35 MPa, pressure drop of 24 MPa, temperature of 50 C, equilibration time of 30 min, collection time of 60 min, extract volume of 30 mL, and feeding solution concentration of 2 mg mL1. Results of 3 replicates (experiments 6, 13, and 14) were indicative of reproducibility of the particle production process. Further studies may be focused on other herbal extracts to determine generalizability of the proposed procedure, and upscaling the procedure, which is one of the perspectives of this research. Acknowledgment The authors have no conflicts of interest to declare. References 1. Cardellina JH. Challenges and opportunities confronting the botanical dietary supplement industry. J Nat Prod. 2002;65:1073-1084. 2. Li SD, Huang L. Pharmacokinetics and biodistribution of nanoparticles. Mol Pharm. 2008;5:496-504. 3. Saraf AS. Applications of novel drug delivery system for herbal formulations. Fitoterapia. 2010;81:680-689. 4. Sinha B, Muller RH, Moschwitzer JP. Bottom-up approaches for preparing drug nanocrystals: formulations and factors affecting particle size. Int J Pharm. 2013;453:126-141.

5. Martín A, Cocero MJ. Micronization processes with supercritical fluids: fundamentals and mechanisms. Adv Drug Deliv Rev. 2008;60:339-350. 6. Reverchon E, Adami R. Nanomaterials and supercritical fluids. J Supercrit Fluids. 2006;37:1-22.     Skerget 7. Knez Z, M, Knez Hrn ci c M, Cu cek D. Particle formation using sub- and supercritical fluids. In: Supercritical Fluid Technology for Energy and Environmental Applications. Amsterdam, Netherlands: Elsevier B.V.; 2014:31-67. 8. Campardelli R, Baldino L, Reverchon E. Supercritical fluids applications in nanomedicine. J Supercrit Fluids. 2015;101:193-214. 9. Fages J, Lochard H, Letourneau JJ, Sauceau M, Rodier E. Particle generation for pharmaceutical applications using supercritical fluid technology. Powder Technol. 2004;141(3):219-226. 10. Debenedetti PG, Tom JW, Kwauk X, Yeo SD. Rapid expansion of supercritical solutions (RESS): fundamentals and applications. Fluid Phase Equilib. 1993;82: 311-321. 11. Helfgen B, Turk M, Schaber K. Theoretical and experimental investigations of the micronization of organic solids by rapid expansion of supercritical solutions. Powder Technol. 2000;110:22-28. 12. Weber M, Thies MC. A simplified and generalized model for the rapid expansion of supercritical solutions. J Supercrit Fluids. 2007;40(3):402-419. 13. Pessi J, Lassila I, Merilainen A, Raikkonen H, Haeggstrom E, Yliruusi J. Controlled expansion of supercritical solution: a robust method to produce pure drug nanoparticles with narrow size-distribution. J Pharm Sci. 2016;105:2293-2297. 14. Byrappa K, Ohara S, Adschiri T. Nanoparticles synthesis using supercritical fluid technologydtowards biomedical applications. Adv Drug Deliv Rev. 2008;60(3): 299-327. 15. Li S, Yuan W, Deng G, Wang P, Yang P, Aggarwal B. Chemical composition and product quality control of turmeric (Curcuma longa L.). Pharm Crops. 2011;2: 28-54. 16. Singh G, Kapoor IP, Singh P, de Heluani CS, de Lampasona MP, Catalan CA. Comparative study of chemical composition and antioxidant activity of fresh and dry rhizomes of turmeric (Curcuma longa Linn.). Food Chem Toxicol. 2010;48:1026-1031. 17. Manzan AC, Toniolo FS, Bredow E, Povh NP. Extraction of essential oil and pigments from Curcuma longa [L.] by steam distillation and extraction with volatile solvents. J Agric Food Chem. 2003;51:6802-6807. 18. Raina VK, Srivastava SK, Syamsundar KV. Rhizome and leaf oil composition of Curcuma longa from the lower himalayan region of Northern India. J Essent Oil Res. 2005;17:556-559. 19. Chang LH, Jong TT, Huang HS, Nien YF, Chang CM. Supercritical carbon dioxide extraction of turmeric oil from Curcuma longa Linn and purification of turmerones. Sep Purif Technol. 2006;47:119-125. 20. Kocaadam B, S¸anlier N. Curcumin, an active component of turmeric (Curcuma longa), and its effects on health. Crit Rev Food Sci Nutr. 2017;57(13):2889-2895. 21. Brumatti LV, Marcuzzi A, Tricarico PM, Zanin V, Girardelli M, Bianco AM. Curcumin and inflammatory bowel disease: potential and limits of innovative treatments. Molecules. 2014;19:21127-21153. 22. Raufa A, Imranb M, Orhanc IE, Bawazeerd S. Health perspectives of a bioactive compound curcumin: a review. Trends Food Sci Technol. 2018;74:33-45. 23. Tang M, Taghibiglou C. The mechanisms of action of curcumin in Alzheimer’s disease. J Alzheimers Dis. 2017;58(4):1003-1016. 24. Xie M, Fan D, Zhao Z, et al. Nano-curcumin prepared via supercritical: improved anti-bacterial, anti-oxidant and anti-cancer efficacy. Int J Pharm. 2015;496(2):732-740. 25. Anand P, Thomas SG, Kunnumakkara AB, et al. Biological activities of curcumin and its analogues (Congeners) made by man and mother nature. Biochem Pharmacol. 2008;76(11):1590-1611. 26. Paulucci VP, Couto RO, Teixeira CCC, Freitas LAP. Optimization of the extraction of curcumin from Curcuma longa rhizomes. Braz J Pharmacogn. 2013;23(1):94-100. 27. Maizura M, Aminah A, Aida W. Total phenolic content and antioxidant activity of kesum (Polygonum minus), ginger (Zingiber officinale) and turmeric (Curcuma longa) extract. Int Food Res. 2011;18:526-531. 28. Martinez-Correa HA, Paula JT, Kayano ACA, et al. Composition and antimalarial activity of extracts of Curcuma longa L. obtained by a combination of extraction processes using supercritical CO2, ethanol and water as solvent. J Supercrit Fluids. 2017;119:122-129. 29. Braga MEM, Leal PF, Carvalho JE, Meireles MAA. Comparison of yield, composition, and antioxidant activity of turmeric (Curcuma longa L.) extracts obtained using various techniques. J Agric Food Chem. 2003;51:6604-6611. 30. Braga MEM, Meireles MAA. Accelerated solvent extraction and fractioned extraction to obtain the Curcuma longa volatile oil and oleoresin. J Food Process Eng. 2007;30:501-521. 31. Ashraf H, Sadiq Butt M, Asghar A, Shahid M. Comparative study of conventional solvent and supercritical fluid extracts of turmeric using high performance liquid chromatography. Pak J Agric Sci. 2016;53(4):941-946. 32. Kwon H-L, Chung M-S. Pilot-scale subcritical solvent extraction of curcuminoids from Curcuma long L. Food Chem. 2015;185:58-64. 33. Chhouk K, Wahyudiono W, Kanda H, Goto M. Comparison of conventional and ultrasound assisted supercritical carbon dioxide extraction of curcumin from turmeric (Curcuma longa L.). Eng J. 2017;21(5):53-65. 34. Wakte PS, Sachin BS, Patil AA, Mohato DM, Band TH, Shinde DB. Optimization of microwave, ultra-sonic and supercritical carbon dioxide assisted extraction techniques for curcumin from Curcuma longa. Sep Purif Technol. 2011;79(1): 50-55.

F. Momenkiaei, F. Raofie / Journal of Pharmaceutical Sciences 108 (2019) 1581-1589 35. Anand P, Kunnumakkara AB, Newman RA, Aggarwal BB. Bioavailability of curcumin: problems and promises. Mol Pharm. 2007;4(6):807-818. 36. Yang KY, Lin LC, Tseng TY, Wang SC, Tsai TH. Oral bioavailability of curcumin in rat and the herbal analysis from Curcuma longa by LC-MS/MS. J Chromatogr B Analyt Technol Biomed Life Sci. 2007;853(1-2):183-189. 37. Araiza-Calahorra A, Akhtar M, Sarkar A. Recent advances in emulsion-based delivery approaches for curcumin: from encapsulation to bioaccessibility. Trends Food Sci Technol. 2018;71:155-169. 38. Mehanny M, Hathout RM, Geneidi AS, Mansour S. Exploring the use of nanocarrier systems to deliver the magical molecule; curcumin and its derivatives. J Control Release. 2016;225:1-30. 39. Zabihi F, Xin N, Jia J, Chen T, Zhao Y. High yield and high loading preparation of curcuminePLGA nanoparticles using a modified supercritical antisolvent technique. Ind Eng Chem Res. 2014;53:6569-6574. 40. Shin GH, Li J, Cho JH, Kim JT, Park HJ. Enhancement of curcumin solubility by phase change from crystalline to amorphous in cur-TPGS nanosuspension. J Food Sci. 2016;81:494-501.

1589

41. Chen C, Sun W, Wang X, Wang Y, Wang P. Rational design of curcumin loaded multifunctional mesoporous silica nanoparticles to enhance the cytotoxicity for targeted and controlled drug release. Mater Sci Eng C. 2018;85:88-96. 42. Jia J, Wang W, Gao Y, Zhao Y. Controlled morphology and size of curcumin using ultrasound in supercritical CO2 antisolvent. Ultrason Sonochem. 2015;27:389-394. 43. Zhao Z, Xie M, Li Y, et al. Formation of curcumin nanoparticles via solutionenhanced dispersion by supercritical CO2. Int J Nanomedicine. 2015;10:3171-3181. 44. Kurniawansyah F, Mammucari R, Foster NR. Polymorphism of curcumin from dense gas antisolvent precipitation. Powder Technol. 2017;305:748-756. 45. Momenkiaei F, Raofie F. Preparation of Silybum marianum seeds extract nanoparticles by supercritical solution expansion. J Supercrit Fluids. 2018;138:46-55.  Jacobsen NE, Timmermann BN, Gang DR. Analysis of cur46. Jiang H, Somogyi A, cuminoids by positive and negative electrospray ionization and tandem mass spectrometry. Rapid Commun Mass Spectrom. 2006;20:1001-1012. 47. Witschi C, Doelker E. Residual solvents in pharmaceutical products: acceptable limits, influences on physicochemical properties, analytical methods and documented values. Eur J Pharm Biopharm. 1997;43:215-242.